AI Takes Centre Stage at MWC 2025: What Comes Next?

March 2025

Telecoms & Connectivity

Molly Gatford

Senior Research Analyst

Ahead of our visit to MWC 2025, I was anticipating lots of announcements and showcases of AI tools this year, and I was certainly not disappointed. Not only was AI+ one of the themes of MWC this year, but many of the exhibitors strapped the word ‘AI’ to their stands, and many of the conversations we had at the event discussed the use of AI solutions for a range of applications.

AI remains a major talking point, and its role in the telecommunications industry is set to grow even further over the next year. At MWC 2025, we saw AI being applied across a wide spectrum - from consumer devices to network infrastructure - alongside the increasing adoption of agentic AI.

In this blog, I’ll share my insights on the AI applications I spotted at MWC and what they mean for the future, starting with:

Agentic AI – What to Expect in 2025

While I was writing our Conversational AI report, published only last month, agentic AI came up in several of the discussions I had with a wide variety of telecoms stakeholders, as a future disruptor of the communications market, so I was excited to hear more at MWC 2025.

Agentic AI has also recently been gaining increased attention within the wider telecommunications industry, with the goal of automating a greater number of tasks. The notable agentic AI announcements made at MWC this year were as follows:

• Nokia announced the use of a telco-trained LLM and agentic AI to monitor the network, to identify threats and retrieve insights. Its new AI-powered Threat Hunt Assistant will leverage telco threat intelligence, network telemetry, and AI to detect attacks and provide guidance to security analysts; significantly reducing the time taken for threat detection. Moreover, Nokia announced that its Digital Operations Center will use agentic AI to automate tasks and troubleshoot issues in service orchestration, fulfilment, and assurance.
• Deutsche Telekom & Google Cloud announced the development of a RAN Guardian Agent, which is a multi-agentic AI assistant built using Gemini 2.0. The agent will monitor the network in real-time and perform self-healing measures to optimise network performance. In turn, this is expected to enhance mobile experience, especially within scenarios with a high network load.
• Honor’s ALPHA PLAN aims to create an open AI ecosystem by bridging the gap between human and AI interaction. The company has announced a $10 billion investment over the next five years to build this ecosystem, with an initial focus on agentic AI. Their first project will be a ‘GUI-based mobile agent’ that can handle tasks like booking restaurant reservations, managing smartphone activities, adding events to a calendar, and more.

It appears that the use of agentic AI within networks will also be set to deliver significant benefits; enabling the automation of a broader range of tasks within network management, security and optimisation. As such, we believe that the term ‘agentic AI’ will continue to grow in popularity over the next year, with more offerings to be developed.

AI in Networks

With the launch of our AI in Networks report forthcoming, I was set to report back on any interesting announcements at MWC 2025. AI can be used in networks to automate and speed up a range of tasks, including network management, optimisation and maintenance. The following announcements of the use of AI in networks were of note:

• China Mobile's newly announced AI+NETWORK initiative focuses on leveraging AI to transform telecommunications networks, enhance 5G-A technology, and create autonomous networks - with a focus on automation, data analytics, and global collaboration.
• KDDI showcased its next-gen AI datacentres that repurpose factories with large power capacity, as well as demonstrations of fault recovery and network construction that use AI.
• SoftBank made several announcements regarding AI in networks, including AI-RAN. It showcased AI for RAN performance enhancement, commercialisation of an AI-RAN solution for centralised RAN and distributed RAN deployments, AI-on-RAN for deploying AI applications on the network edge, and the optimisation of resource allocation for AI applications.

These announcements highlight that the Asia Pacific region is at the forefront of implementing AI in networks, with telcos in the region driving innovation. This will likely lead to better network monetisation in the coming years. We anticipate that telcos in other regions will follow suit, adopting AI to boost automation and optimise network resources through dynamic allocation, while proactively identifying network failures to minimise downtime and secure ROI from 5G networks.

Molly is a Senior Research Analyst at Juniper Research, providing insight, data, and recommendations for established and future markets within the telecommunications sector. Her recent reports have covered topics including Conversational AI, RCS Business Messaging, and AdTech.